1 Introduction
Since 2004, graphene-based materials (GBMs), which are defined as carbon-based 2D structures, have gained a lot of attention because of their exceptional properties on, for instance, electrical and thermal conductivity (Brownson et al.
2012; Novoselov et al.
2004). Hence, researchers, companies, and governments around the world have recognised that these carbon-based nanomaterials are key contenders to improve several modern devices and to offer new solutions in high-tech sectors such as photonics (Li et al.
2019; Sankar et al.
2019; Ye and Tour
2019). The European commission has also shown high expectations for GBMs with significant financial support for their strategic development (e.g. funding of the “Graphene Flagship” project). Furthermore, experts around the world expect a rise of their use within the next decades (Geim
2009; Novoselov et al.
2012; Randviir et al.
2014). With such an outlook on the future of GBMs, their environmental impacts should be well understood to evaluate if their integration into our everyday life might help in building a more sustainable future. The life cycle assessment (LCA) methodology (ISO14040
2006; ISO14044
2006; Joint Research Center
2010) is a suitable choice to perform such assessments (Ness et al.
2007; Rebitzer et al.
2004), and it has been previously used in the sector of nanomaterials (Salieri et al.
2018).
In 2017, Arvidsson reviewed four LCA studies (Arvidsson et al.
2016,
2014; Arvidsson and Molander
2016; Pizza et al.
2014) on this subject and highlighted key sources of environmental impacts for five production methods. Concurrently, other publications (Cossutta et al.
2017,
2020; Scott and Cullen
2016) provided similar analyses of diverse GBMs’ production methods. Table
1 presents an overview of these studies and the identified main sources of environmental impacts, which highlights some discrepancies for similar production processes (e.g. chemical reduction). The explanation for these divergent observations could be linked to the different input flows (substances and energy sources) or the differences in modelling assumptions, but the answer is not yet identified. Moreover, the use of diverse modelling assumptions in different studies impedes the comparison of quantified environmental impacts between production processes. Indeed, the use of different databases and impact assessment methods in the studies of Table
1 precludes any direct comparison of available results. A compatible model (i.e. with similar modelling assumptions) for all production processes, which has not been fully provided in previous publications, could therefore build on the current knowledge to identify key aspects for the development of GBMs. It could also help in the improvement of production processes to identify the best development options from an environmental perspective.
Table 1
Main sources of potential environmental impacts for seven production processes of graphene-based materials (GBMs)
Chemical reduction | Reduction step (i.e. Hummers’ process) | |
Chemical reduction | Energy demand (i.e. electricity) | |
Thermal reduction | Energy demand (i.e. electricity) | |
Exfoliation by ultrasonication | Chemical substance (i.e. diethyl ether) | |
Electrochemical exfoliation | Energy demand (i.e. electricity) | |
Thermal exfoliation | Energy demand (i.e. microwave heating) | |
Chemical vapour deposition | Carbon feedstock (i.e. methane) | |
Chemical vapour deposition | Carbon feedstock (i.e. methane) | |
Chemical vapour deposition | Energy demand (i.e. electricity) | |
Epitaxial growth | Carbon feedstock (i.e. silicon carbide) | Arvidsson and Molander ( 2016) |
Another challenge in the analysis of previous studies comes from the lack of measurements for the functional properties (e.g. conductivity) of produced GBMs and their different states (solution or dry mass). This observation has been made by Arvidsson (
2017), but a quantified evaluation of their influences on results has not been found in recent publications. Nevertheless, the use of GBMs has been considered for specific applications like: additive for polymers (Pizza et al.
2014) and coatings (Upadhyayula et al.
2017), transparent conductive layers (Arvidsson et al.
2016), or photovoltaic back-contact material (Scott and Cullen
2016; Scott et al.
2016). These examples shed some light on the relevant properties to define the functionality of GBMs.
The frequent use of theoretical values from scientific publications and patents is another important aspect of recent LCA studies on GBMs. It shows that information from GBM producers is still difficult to obtain, but it has also been suggested as a relevant proxy for industrial-scale production when production efficiency could be optimised and when GBMs should be economically competitive. Some studies on future industrial-scale production have not, however, considered the possible evolution in background systems (Arvidsson et al.
2016,
2014) while others have used electricity mixes of specific countries to represent best- and worst-case scenarios (Arvidsson and Molander
2016; Cossutta et al.
2017; Pizza et al.
2014). This choice of presenting best- and worst-case results is understandable, but these analyses could benefit from a wider range of forecasts if decarbonised options are not available to some producers. Furthermore, four publications have used version 2.2 of the ecoinvent database (Arvidsson et al.
2016,
2014; Arvidsson and Molander
2016; Pizza et al.
2014), thus reducing the temporal representativeness of the modelled electricity mixes. Indeed, the model of the European electricity mix in version 2.2 is based on statistics of 2000 whereas version 3.4 uses electricity market statistics from 2014.
Tackling the limitations of past studies now seems essential to build on the knowledge that has been offered by previous assessments and to provide general guidance on the potential improvement of environmental sustainability for future production of GBMs. This work thus uses a harmonised modelling framework to assess the environmental impacts of production processes from Table
1. Key options for improvements and processes with overall lower environmental impacts can then be identified with this broad assessment. New lab-scale data, which has been measured in situ, is also used in order to shed some light on the variations of impacts when GBMs are offered in different states. These new results are then compared with the values of the studies listed in Table
1. A subsequent assessment of uncertainty and variability for all these production processes then provides the range of impacts that can be derived from current knowledge. Lastly, prospective models for electricity production are applied to assess the potential future variation of impacts for GBMs and competing materials. These results are then used to quantify the improvement in efficiency that GBM production methods should reach to offer environmentally competitive options with some substances that they might replace. Detailed information on all steps of this analysis is provided in
supplementary documents for the scope definition (
SISD), datasets’ descriptions (
SIID), and results (
SIR).
5 Discussion
The first general observation that comes out of this study is the key roles that chemical substances and electricity demand play in the impacts of a GBM’s life cycle. Their importance varies depending on the type of GBM and the production process. For instance, potential environmental impacts from GO production mainly depend on chemical substances, while impacts for oGBM production depend on its electricity demand except for the ultrasonication process, which is dominated by the use of diethyl ether. These results confirm most conclusions from previous publications (in Table
1) and highlight key inputs on which the industry should focus for a more efficient reduction of environmental impacts in future GBM production processes. Two recommendations thus come to mind:
Such recommendations are compliant with the 12 principles of green chemistry,
1 which provide a good general strategy to reduce the environmental impacts of GBM production.
The contribution analysis shows that most of the identified discrepancies between studies can be explained by the difference in aggregation levels of the various system models rather than by actual differences in inputs for the considered production methods. This observation shows the relevance of an analysis with a harmonised framework, since it clarifies that previous studies often show agreement on key contributors of environmental impacts from GBM production.
The use of conservative prospective scenarios from the IEA for the European electricity mix then shows that possible evolution pathways until 2040 would not often affect the ranking between production processes of GBMs even if electricity demand is a key contributor for some production options. This interpretation provides a different perspective than the previous sensitivity analyses of Arvidsson et al. (
2016,
2014) and Cossutta et al. (
2017) where best-case scenarios were presented with highly decarbonised electricity sources (e.g. ~ 30 g of CO
2 eq./kWh). The use of three scenarios for 2025, 2030, and 2040 nevertheless shows some changes in the ranking between the carbon footprint of CRR, TRR, and ECE production options, mainly for the SDS 2040 forecast. This suggests that electricity inputs with a carbon footprint below ~ 110 g of CO
2 eq./kWh could be relevant for the environmental competitiveness of ECE production.
The evaluated impacts for the purified solution or dry mass from the BAM process offer new insights on the importance of the output state for GBMs. Indeed, keeping the GBM in a solution might be environmentally beneficial if it can be used directly as an input for the production of devices. These results also hint that recycling of chemical substances might come with a significant increase in environmental impacts for GBM products if they require dialysis of their solutions.
The combination of the broad analysis on energy use and the new data on BAM production thus provides new insights:
-
The choice of specific energy sources or country of production could bring significant decrease in impacts and might have more effects than the IEA forecasted evolution pathways of the European electricity mix.
-
Energy efficient recycling options to reduce the use of chemical substances might be critical to ensure environmental improvements of GBMs.
-
Choosing the most relevant state of GBM (i.e. in solution or in dry mass) for their use in the production of different devices should have substantial effect on the impacts of such devices.
The evaluation of uncertainty for all GBM production pathways indicates that there is still a considerable ambiguity on their potential environmental impacts except for the most recent studies (Cossutta et al.
2017,
2020). Indeed, the oldest publication on production of GBM (Arvidsson et al.
2014) shows much wider potential variability of impacts that might be explained by the exploratory nature of the work. The levels of uncertainty on production methods for most GBM types thus prevent the identification of a clear production winner from the environmental perspective. When results from the comparison of impacts and the contribution analysis are combined, electricity demand comes up as the main source of uncertainty in current models, even if it is not the main contributor for all processes. This is explained by the substantial level of variability that can be observed for the inputs of electricity demand in all studies and the unknown nature of electricity production when GBM manufacturing might reach industrial levels.
The comparison of GR with other competing materials (i.e. graphite, molybdenum, ITO) suggests that some production options might result in a decrease of the environmental impacts of touch screens. Conversely, graphite is often a better environmental choice for the case of back contacts in photovoltaic cells. It is worth mentioning that these conclusions are still preliminary since substantial uncertainties are afflicting the comparisons. The only clear result is that the EPI production option will not offer environmentally competitive GR unless electricity demand can be reduced by more than ~ 99.99% when the process is industrialised.
While this study provides some new suggestions to decrease the environmental impacts of GBM production, it is also limited on some aspects that should be tackled in future LCA studies if more information becomes available.
Indeed, quantified descriptions of functional properties for all GBM are still missing, which hinders from making balanced comparisons of GBMs for their specific uses in devices. In this study, it mainly affects the extent of the analysis when oGBMs are compared because it is unlikely that they all have equivalent functional properties that would make them interchangeable in potential future devices. The PAS 1201:2018 (BSI
2018) document provides recommendations and guidance on the properties that should be reported for different GBMs, which could constitute a good starting point to identify the relevant functional properties for future LCA studies.
Additionally, the measured electricity demand for BAM production raises questions on ranking with previous estimations since there are significant uncertainties on values from some past studies that have used patents and scientific articles instead of measures (Arvidsson et al.
2016,
2014; Arvidsson and Molander
2016; Pizza et al.
2014; Scott and Cullen
2016). The different ranges might be explained by the use of different equipment or the lab-scale production of BAM or the purity/quality of produced oGBMs, but identifying the key sources of discrepancies will be challenging until further investigations are made on all processes.
Reducing uncertainties on all input data should also be a priority to identify if the life cycle impacts of GBM production processes are still overlapping or if some options might have overall better environmental performances. This is a more pressing matter for GO and oGBM production options even if GR uncertainties prevent from identifying any distinction between two types of CVD production processes (i.e. CVD 3 and CVD 4). Measurements of inputs flows at the lab-scale production would be a good way to reduce these uncertainties until industrial scale production is reached.
The toxicity of GBMs is currently evaluated within the European Graphene Flagship research initiative (Fadeel et al.
2018), and some characterisation factors (CFs) for GO have been proposed for freshwater (Deng et al.
2017). Nevertheless, more research will be necessary to evaluate the potential impacts of all GBMs within the LCA framework. For now, the baseline CF, provided by Deng et al. (
2017) (i.e. 777.5 CTU
h·m
3·year/kg of GO), shows that GO emissions in freshwater might bring significant change on the FEcotox impact. Indeed, the FEcotox impact of the H5 production process would double if 41% of the produced GO were released, for instance, at its end of life. On the other hand, this impact of GO would be less of an issue for production processes with higher impacts (e.g. H3) where a 100% release would only increase its FEcotox impact by ~ 4%.
When functional properties will be defined for all GBMs, models for the use and end-of-life phases should be explored because it will then be possible to assess the amount of GBM that is necessary for different devices. The use of new CFs should also be important to fully consider the potential environmental impacts of GBMs. Change in the lifetime of devices, when they use GBMs instead of competing material, might then become a relevant aspect to add in the assessment.
6 Conclusion
The investigation of prior LCA studies and new data for GBM production, within a harmonised modelling framework, provides additional insights on the origins and significance of environmental impacts from all options for GBM production. It also shows how various impacts might be reduced while reinforcing some of the conclusions from previous studies when common observations are reached.
Overall, agreements are found on the necessity to focus on the reduction of impacts from chemicals and electricity uses in GBM production processes to improve the environmental sustainability of their use in future devices. Attention to different aspects of production depends on the type of GBM and its output state (e.g. in solution or dry mass). Concerning energy use, using only renewable energy sources would be a relevant choice to substantially reduce the impacts for some production methods. On the other hand, a conservative evolution of the European electricity mixes until 2040 might not bring many important changes in the current environmental ranking of GBMs. Furthermore, while some production methods show significantly lower environmental impacts than others, comparisons with competing materials also illustrate that GR might not always be the best environmental option for some uses. Finally, the analysis of previous publications shows that industrial-scale production could reduce the environmental impacts of production.
Many aspects that affect the environmental performances of GBMs are still not well described or understood, which raise important questions on the current assessments. Indeed, quantified evaluations of functional properties and toxicity of GBMs are good examples of unresolved questions that would benefit from further investigations. Precise values on the energy use during the production of GBMs would also be beneficial to increase our capacity to differentiation current and future production options.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.